Chemico-Biological Interactions 192 (2011) 129–135

Contents lists available at ScienceDirect

Chemico-Biological Interactions
journal homepage: www.elsevier.com/locate/chembioint

Proteomics of toxic oil syndrome in humans: Phenotype distribution in a

population of patients
Carmen Quero b , Nuria Colomé a , Carlos Rodriguez a , Peter Eichhorn a , Manuel Posada de la Paz c , Emilio
Gelpi a,∗ , Joaquin Abian a
a
CSIC-UAB Proteomics Laboratory, IIBB-CSIC/IDIBAPS, Barcelona, Spain
b
Department Química Biológica IQAC, CSIC, Barcelona, Spain
c
CISATER, Instituto de Salud Carlos III, Madrid, Spain

a r t i c l e i n f o a b s t r a c t

Article history: Objectives: Toxic oil syndrome (TOS) is a disease that appeared in Spain in 1981. Epidemiological work
Available online 12 November 2010 traced the origin to the ingestion of aniline-adulterated rapeseed oil, fraudulently marketed and sold as
edible oil. It affected more than 20,000 people with over 400 deaths in the first 2 years. In 2001 evidence
Keywords: was presented that genetic factors could play a role in the susceptibility of individuals to the disease. Thus,
Toxic oil syndrome a prospective study on the differences in gene expression in sera between control versus TOS-affected
Disease markers
populations, both originally exposed to the toxic oil, was undertaken in our laboratory.
Haptoglobin polymorphism
Methods: Differential protein expression was analyzed by two-dimensional electrophoresis (2-DE). Prob-
Serum proteomics
MALDI-TOF proteotyping
lems related with serum analysis by 2-DE were addressed to improve protein detection in the gel images.
Three new commercial systems for albumin depletion were tested to optimize the detection of minor
proteins. The use of nonionic reductants or the presence of thiourea in the gels, were also tested.
Results: From the resulting optimized images, a group of 329 major gel spots was located, matched
and compared with serum samples. Thirty-five of these protein spots were found to be under- or over-
expressed in TOS patients (threefold increase or decrease). Proteins in these spots were identified by
matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) peptide map fingerprinting and
database search. Several haptoglobin (Hp) isoforms were found to be differentially expressed, showing
expression phenotypes that could be related with TOS. Resolution of the homologous ␣-1s and ␣-1f
chains, with a mass difference of only 0.043 Da, was obtained after guanidation of the protein with O-
methylisourea. We applied this procedure to the study of the distribution of the Hp alleles HP2 , HP1s and
HP1f in control versus TOS-affected populations. The MALDI-TOF proteotyping method was validated by
a parallel analysis of the serum samples by 2-DE.
Conclusions: Data obtained from 54 TOS cases and 48 controls indicate significant differences in the
distribution of Hp phenotypes in the two populations. Haptoglobin phenotypes have been reported to
have biological and clinical consequences and have been described as risk factors for several diseases.
Consequently, it was concluded that haptoglobin polymorphism could play a role in TOS.
© 2010 Elsevier Ireland Ltd. All rights reserved.

1. Introduction cough, dyspnea, pulmonary infiltrates, severe muscle pain, intense

The toxic oil syndrome (TOS) appeared in Spain in 1981 due
to the ingestion of an illegally commercialized aniline-denatured
rapeseed oil. This oil, imported for industrial uses, was refined
and mixed with other oils to be sold as edible oil. TOS pre-
sented a complex symptomatology with different phases including
∗ Corresponding author at: Instituto de Investigaciones Biomédicas de Barcelona
(IIBB-CSIC-IDIBAPS), Laboratorio de Proteómica CSIC/UAB, Facultad de Medicina
(UAB), Roselló 161, 6a Planta, E-08036 Barcelona, Spain. Tel.: +34 933638300x345;
fax: +34 933638301; mobile: +34 639121499.
E-mail addresses: egmbam@iibb.csic.es, egelpi@gmail.com (E. Gelpi).
eosinophilia, ascending polyneuropathy, sclerodermiform skin
changes and pulmonary hypertension [1]. TOS affected more than
20,000 people, with over 300 deaths in the first two years. About
2500 deaths were related to the disease until 1998 and there
are people still affected with various degrees of disabilities [2].
Although several chemical markers of the toxic oil are known
(anilides, phenylaminopropanediols) [3], the actual toxic factor(s)
remain obscure. A further understanding of those factors is of prime
importance to prevent the occurrence of any other related episode,
such as the eosinophilia myalgia syndrome that appeared in 1989
in the USA [4].
The contribution of genetic factors to the risk of suffering the
disease after oil consumption was suggested [5]. Thus, we carried

0009-2797/$ – see front matter © 2010 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.cbi.2010.11.001
130 C. Quero et al. / Chemico-Biological Interactions 192 (2011) 129–135
out a prospective study on the differences of gene expression in
sera between control versus TOS-affected populations, both origi-
nally exposed to the toxic oil. Preliminary work from pools of sera
showed that two haptoglobin (Hp) spots identified as the ␣-1s and
␣-1f chains showed a significantly different distribution in the two
populations studied. This indicated that Hp phenotypes could be
related to individual susceptibility towards the toxic oil syndrome
(TOS) [6]. Hp belongs to the group of the acute-phase proteins
and in addition to its hemoglobin binding function, modulates the
response to infection and inflammation, and is involved in immune
suppression. Human Hp is a tetramer composed of two beta (or
heavy) and two alpha (or light) chains connected by disulfide bonds.
The beta subunits always have the same structure [7] but there
are three forms of the alpha subunit (␣-2, ␣-1s and ␣-1f) coded
by three Hp alleles (HP2 , HP1s and HP1f [8]). The combinations
of these main Hp alleles produce six major expression pheno-
types (Hp2-2, Hp2-1s, Hp2-1f, Hp1s-1s, Hp1s-1f, and Hp1f-1f).
Haptoglobin phenotypes have been reported to have important bio-
logical and clinical consequences and have been described as risk
factors for several diseases. The possession of a particular pheno-
type has been associated with the prevalence and clinical evolution
of many inflammatory diseases, including infections, atheroscle-
rosis, cardiovascular diseases and autoimmune disorders. Many
studies have pointed out that TOS is a disease directly related with
the immune system and with some characteristics similar to those
observed in autoimmune disorders [9,10]. The HP2 allele has been
reported as a factor producing a higher immune reactivity when
compared with the HP1 allele. These facts could point to a higher
response to the oil xenobiotics in individuals bearing the HP2 allele
resulting in a more effective protection.
Analysis of differential proteins in blood or sera has been exten-
sively reported in studies of renal and hepatic failure, malnutrition,
cancer, AIDS, diabetes, and cardiovascular and neurological dis-
eases [11–14]. Proteomic technologies can provide libraries of
biomarkers and immunoassay screens for high-throughput mon-
itoring of exposure to xenobiotics. A major drawback of protein
analysis in blood or sera derives from the presence of a few fam-
ilies of high abundance proteins that can hinder the detection of
other minor proteins in the sample or impede their quantitation.
Thus, the specific removal of albumin (71% of the total protein con-
tent [15]) and other major proteins from serum samples can be
primordial for proper analysis of global protein expression. The use
of antibodies appears to be the most effective way to deplete major
blood proteins [16]. Thus, at the time we did our initial work [6]
we compared several commercial kits for their depletion capacity
and specificity. Although there are available nowadays several com-
mercial kits capable of removing more than 20 of the major serum
proteins families, the situation was not so good at the time we
developed our protocols so that we had to check the few depletion
kits available at that time for their depletion capacity and speci-
ficity. These limitations had, however, a positive outcome, as for
practically all the kits currently in use, haptoglobins are now one of
the main targets for removal after albumin and immunoglobulins.
Other factors affecting the final gel images such as the concen-
tration of chaotropes in the rehydration buffer, the sonication of
the sample, as well as the use of antioxidants such as hydrox-
yethyl disulfide (HED) and dithiothreitol (DTT) were also studied.
Optimized conditions were applied to the search of differentially
expressed proteins in control and disease individuals that could be
used for disease markers in a population.
To circumvent the technical and economical limitations of
the gel-based approaches, we also developed new direct MALDI-
TOF profiling procedures for Hp subtyping that allows the rapid,
automatable analysis of a large number of samples [17,18]. We val-
idated the method in sera from a small group of TOS controls (n = 54)
and cases (n = 48), which were analyzed in parallel by minigel 2-DE.
2. Materials and methods
The equipment for 2-DE and image analysis was from Amer-
sham Biosciences (Uppsala, Sweden). Immobiline DryStrips and
other reagents were also from Amersham and Sigma (see Ref. [6] for
details). The mass spectrometric experiments were carried out in a
Voyager DE-PRO MALDI-TOF mass spectrometer (Applied Biosys-
tems).
2.1. Human serum samples
Serum samples were obtained from the CISATER (Research Cen-
ter for the Toxic Oil and Rare Diseases, Madrid, Spain). Control
samples were from members of affected families, who at the time
of the epidemic, lived in the same house as the rest of the fam-
ily, had regular meals with them and who, despite this, did not
develop TOS. Samples from TOS-affected individuals were obtained
from those patients chronically ill during the first year of the dis-
ease. Several control and disease groups of 4–6 individuals each
were created on the basis of homogeneity of sex and age. A total
of five different pools, corresponding to 50 individuals (15 control
males, 15 patient males, 10 control females and 10 patient females)
were analyzed. Aliquots of 25 ␮L serum from each individual in a
group, were pooled and then aliquoted in 25 ␮L group replicates.
The samples were stored at −80 ◦ C until analysis. Gel separation
and sample treatment were optimized using a pool of sera from
healthy individuals obtained from the CISATER and Vall d’Hebron
Hospital (Barcelona). Serum samples for proteotyping studies were
obtained from a total of 54 chronically ill patients, together with 48
control individuals (see above). These groups included male and
female individuals with an age range from 10 to 70 years. The pro-
portions of males were 65% and 75% in control and case populations,
respectively. Samples were stored at −80 ◦ C.
2.2. Peptide mass fingerprinting (PMF) analysis of gel spots
Hp chain spots were excised and subjected to in-gel digestion
with trypsin, as described by Shevchenko et al. [19]. Tryptic pep-
tides were extracted from the gel with ACN/0.5% TFA 1:1 and the
extracts were evaporated to dryness and redissolved in ACN/0.1%
TFA 1:1. PMF was performed in the Voyager DE-PRO mass spec-
trometer. A small aliquot (0.5 mL) of the tryptic extract was mixed
with 0.5 ␮L CHCA (5 mg/mL).
External calibration was carried out with a set of synthetic pep-
tides. When peptides from autolytic degradation of trypsin were
detected, an internal calibration of the spectra was performed.
2.3. Proteotyping by MALDI-TOF profiling
Serum samples (8 ␮L) were mixed with 2 ␮L 1 M DTT and were
left for 30 min at room temperature to reduce the disulfide bonds
between the Hp␣-1 and b chains. One microliter of the mixture was
diluted 50-fold with water and mixed vigorously. An aliquot (1 ␮L)
was loaded onto the MALDI plate followed by 1 ␮L of the sinapinic
acid matrix (10 mg/mL in ACN/TFA 0.1%, 7:3). The MS analyses were
accomplished in the Voyager DEPRO MALDI-TOF mass spectrom-
eter working in the linear positive ion mode at 25 kV acceleration
voltage. Hp␣-1 and b-2 chains appeared in the spectra at m/z 9192
and 15946, respectively [17].
2.4. Identification of ˛-1s and ˛-1f chains in individuals with 2-1
or 1-1 phenotypes
Target sera (2 ␮L) were mixed with 18 ␮L 2 M O-methylisourea.
The pH was increased to 10.5 with NH4 OH and the mixture was
 C. Quero et al. / Chemico-Biological Interactions 192 (2011) 129–135
incubated for 24 h at 30 ◦ C. After guanidination, NH4 OH was evap-
orated (30 min uncapped) and 1 ␮L 1 M DTT was added to 9 ␮L of
the evaporated solution, which was incubated again for 30 min at
room temperature. After diluting this solution fivefold with water,
the sample was desalted using a C18 micropipette ZipTip. The Zip-
Tip was activated with ACN/0.5% TFA 1:1 (10 ␮L, 5×) and washed
with 0.1% TFA (10 ␮L, 2×). The sample was loaded, then washed
with 0.1% TFA (10 mL, 106) and eluted with ACN/0.5% TFA (4:1)
(10 mL, 10×). The eluted sample was evaporated to a final volume
of about 3 mL, and then was loaded onto the MALDI-TOF plate. After
partial evaporation of the droplets at room temperature (20–25 ◦ C),
1 ␮L of the sinapinic acid matrix was added. MALDI-TOF spectra
were acquired as indicated above. The presence of the ␣-1s and
␣-1f chains was determined by the presence of peaks at approxi-
mately m/z 9528 (8 lysine residues modified) and 9570 (9 residues
modified), respectively [17]. The procedure described above (Sec-
tions 2.3 and 2.4) has been the object of a recent patent entitled
“Procedure for the phenotyping of haptoglobin in human serum”
[18].
2.5. Albumin depletion and 2-DE
Three different kits were tested for their efficiency on albu-
min removal as well as the quality of the 2-DE pattern obtained:
Aurum kit (Bio-Rad), the Montage Albumin depletion kit (Millipore)
and the POROS anti-HSA (Applied Biosystems). The method for the
POROS anti-HSA was optimized for our special requirements, as
described [6]. Once the serum was depleted, the sample was made
up to 350 ␮L with rehydration buffer containing 2% CHAPS, 1% IPG
buffer, 8 M urea, 60 mM DTT, and 100 ␮g/mL of bromophenol blue.
The effect of thiourea and non-ionic reducing buffers in the separa-
tion was also tested. Protein spots in gels were visualized by silver
staining using conventional protocols [19].
2.6. Computer analysis of 2-DE patterns and MS protein
identification
Gel images were digitized using an Agfa DueScan scanner. Pro-
tein patterns were analyzed using Image Master 2D v.4.01c. Protein
spots were characterized with respect to their Mr and pI with 2-D
SDS-PAGE standards. Mr and pI values were automatically calcu-
lated for the remaining protein spots by the Image Master software.
The spots of interest were excised and subjected to in-gel digestion
with trypsin as described by Shevchenko et al. [19].
3. Results and discussion
The number of spots detected as well as the resolution of the
gel images could be improved using nonlinear IPG strips for the
first dimension as well as a gradient gel for PAGE. However, to
obtain a robust and reproducible method for image analysis of
a high number of gels, we choose to use custom-made 12% gels
for the second dimension. The use of nonlinear strips for the first
dimension did not show, in our hands, special advantages for global
analysis over linear strips. Several other procedures for sample
treatment and separation were also tested in order to optimize
the gel images. The results obtained with sonication and albumin
depletion led us to select the POROS system. Various combinations
of urea/thiourea and reducing agents (HED/DTT) were also tested
and already reported [6].
3.1. Analysis of serum samples from control and TOS-affected
individuals
The serum protein profiles from control and TOS patients were
compared with the aim of finding differentially expressed proteins
131
Fig. 1. 2-DE map showing all proteins analyzed by MALDITOF.
Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission from
Ref. [6].
that could be related to organ toxicity. The analysis was initially
carried out from pools of sera from 4 to 5 individuals that were
created on the basis of homogeneity of sex and age. Pool analy-
sis was performed in order to smooth individual differences and
to enhance characteristic traits related with the common disease
affecting these individuals. 6 ␮L of a pooled serum sample was
diluted in PBS, sonicated and treated to eliminate albumin using the
POROS anti-HSA. The extracts were precipitated with TCA/acetone
and the re-dissolved pellet was analyzed by 2-DE. A representative
2-DE gel is shown in Fig. 1. Four image gels of each treatment were
averaged and set for comparison. Only proteins present in at least
two gels in each treatment were considered. A group of 329 major
gel spots was located, matched and compared. Thirty-five of these
protein spots were found to be under- or overexpressed in TOS
patients (twofold increase or decrease). Another 70 spots could not
be matched between gels. Unmatched spots probably correspond
to de novo produced species or to proteins expressed in very low
copies in one of the samples compared. Although a global expres-
sion analysis was intended, special attention was paid to low Mr
molecules because these proteins have been often associated with
a wide range of pathological states such as cancer, AIDS, diabetes,
cardiovascular and neurological diseases [11–14]. One of the most
interesting and clear differences between groups was found in this
area and involved spot nos. 285–287 and 315–316 (Fig. 1). These
spots were identified by MALDI-TOF PMF as different isoforms of
haptoglobin alpha chain. Assignation of each isoform could be done
on the basis of its position in the gel [20] as well as from specific
tryptic peptide signals observed in the MALDI spectra. In this way,
spot nos. 315 and 316 were identified as the f and s isoforms of
haptoglobin alpha 1-chain, respectively, and spot nos. 285, 286 and
287 were identified as different isoforms of the alpha 2-chain. The
image analysis showed that spot nos. 315 and 316 were differen-
tially expressed in control and case samples (Student’s t-test, p,
0.05). The normalized volume of spot 315 was 6.7 ± 5.1 in control
samples and 1.6 ± 0.72 in case samples, for spot 316 the volume was
1.4 ± 0.81 and 21.5 ± 8.4 for control and case samples, respectively.
This behaviour was confirmed in 10 pools of control and case sera
including 50 individuals and covering an age range from 11 to 60
years old male and female pools (Fig. 2, left panel).
Comparison of the theoretical phenotypes in Fig. 2 (right panel)
with the images obtained from control and case sera (Fig. 2, left
panel) indicates, as expected from the use of pools of sera, that these
132 C. Quero et al. / Chemico-Biological Interactions 192 (2011) 129–135
Fig. 2. (Left panel) Comparison of haptoglobin phenotypes in different pools of control and case sera. Pool number is indicated in each figure. (Right panel) Schematic
representation of the different phenotypes derived from haptoglobin alpha 1- and alpha 2-chain expression. TTR, transthyretin; f and s, alpha 1-chain isoforms (white); a2
and other white spots, alpha 2-chain isoforms.
Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission from Ref. [6].
images result from the mixture of individual phenotypical images.
In the case of control pools, the average phenotype observed would
correspond to a mixture of homozygote or heterozygote genotypes
for alpha 1- and 2-chains, the Hp2-2 phenotype probably being the
most frequent one. On the contrary, case pools show profiles that
would agree with phenotypes Hp2-1s, Hp1s-1s or its mixture.
With these results in hand a more detailed study was under-
taken on individual samples to clarify the role of haptoglobin
phenotypes as a marker of risk factor for this disease as well as
the mechanisms implicated in its etiology.
3.2. Hp direct typing by MALDI-TOF
The high abundance of the Hp molecules and the relatively low
molecular weights of its chains made the MALDI-TOF approach a
good candidate for their direct monitoring in serum samples. To test
this approach, we analyzed standard Hp samples and untreated,
diluted human sera from a population of 54 TOS cases and 48 con-
trol individuals by linear MALDI-TOF using sinapinic acid as matrix
[17]. This method, which we have recently patented [18], avoids
complex sample preparation procedures as well as the prior elec-
trophoretic separation of the proteins in the samples and facilitates
the rapid subtyping of haptoglobin. As shown in Fig. 3, Hp␣-1 and
␣-2 chains were easily detected in the MALDI-TOF spectra obtained
from DTT-treated and diluted serum samples. Reduction of the
sample was required to separate the individual chains of the Hp
tetramer that were covalently linked by interchain disulfide bonds.
In these spectra, the ␣-2 chain appeared as a mass signal cen-
tered at about m/z 15945, while the ␣-1s and ␣-1f chains appeared
unresolved at m/z 9193 (external mass calibration). The expected
molecular masses calculated from their amino acid sequences were
15946 and 9192, respectively. Identification of these signals in the
spectra was performed by spiking serum samples with a Hp stan-
dard that contained a mixture of the polymorphic forms of the
protein. Identification was also validated with a group of serum
samples previously phenotyped by 2-DE (Fig. 1). Despite the fact
that in this way haptoglobin typing can be carried out directly on
unpurified serum samples, possible interferences can further be
minimized by in situ purification on the MALDI plates with immo-
bilized Hp antibodies [18]. Although this approach has not been
implemented in this case for the number of samples analyzed,
these immuno-MS methods should be taken into consideration as
they have being shown to provide the robustness and sensitivity
required in large scale, robotized analyses were a high number of
different samples with minimal pre-treatment have to be analyzed
[21,22]. Thus an immuno-MS method was specifically developed
and patented for fast haptoglobin genotyping in serum [16], as
schematically illustrated in Fig. 4.
Direct MALDI analysis of the serum samples did not, however,
discriminate between the ␣-1s and ␣-1f chains. The amino acid
 C. Quero et al. / Chemico-Biological Interactions 192 (2011) 129–135 133

Fig. 3. MALDI-TOF profiles obtained from (A) a Hp standard treated with 5 mM DTT, and (B) a pool of sera treated with 0.2 M DTT. The signals from the ␣-1 and ␣-2 chains
of the Hp molecule are observed at about m/z 9192 and 15946 Da, respectively.
Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission from Ref. [17].

sequence of these chains differs only at positions 52 and 53, where
the pair Asn-Glu replaces Asp-Lys. This produces a mass difference
of only 0.043 Da between the two chains. The sequence of the ␣-
1s chain contains 8 Lys residues, while an additional Lys is present
in the non-homologous amino acids pair of the ␣-1f chain. This
difference can be exploited through the chemical modification of
the Lys residues to produce protein derivatives of different masses
that could be resolved in the spectra. Conversion of the Lys residues
into homoarginine by guanidination at the e-amino group of the
lysine [23] results in a mass difference of 42 Da between the fully
modified ␣-1s and ␣-1f chains. Thus, after guanidination, the signal
of the unmodified chains at m/z 9192 is shifted by 8 (␣-1s) or 9 (␣-
1f) multiples of 42 Da, resulting in two possible signals appearing
at about m/z 9528 and 9570, respectively, which allow the specific
identification of the Hp subtypes (Fig. 5).
To validate the method, the samples were analyzed in paral-
lel by 2-DE. The procedure was similar to that reported above [6]
but in this case, 7-cm strips and minigels were used to speed up
the analysis. The percentage of acrylamide and other 2-DE condi-
tions was optimized to obtain clear images of the Hp ␣-chain region
that would allow rapid evaluation of the Hp-derived spots. In these
conditions, removal of major serum proteins such as albumin or
immunoglobulins was not necessary as they did not interfere with
the target area of the gel image. Among the 102 samples analyzed
by 2-DE, 4 could not be phenotyped due to the complexity of the
spot profiles. Additional interfering spots observed in these sam-
ples could be either genuine components of the proteome or the
result of inadequate storage or handling of the serum sample and
consequent protein degradation. The presence of these modified
forms was the main limitation of the 2-DE approach for Hp phe-
notyping. These samples were readily identified by MALDI-TOF,
thus confirming their previous assignment by 2-DE and PMF. The
presence of modified forms did not interfere with the mass anal-
ysis as the different species produced derivatives with different
masses, thus allowing an unambiguous identification. The confi-
dence of the MALDI-TOF data was also shown in the case of two
mismatched samples that were differently assigned by 2-DE and
MALDI. A reanalysis by 2-DE demonstrated that these mismatches
were the result of the lack of clearly visible Hp spots at the corre-
sponding location and to the presence of faint spots that produced
false-positive identifications, respectively. Despite the problems
encountered in the 2-DE analysis, this sample was readily iden-
tified as Hp2-1s by MALDI-TOF prototyping, indicating the higher
sensitivity and specificity of the latter approach.

Fig. 4. Schematic representation of the optional purification of serum samples directly on MALDI plates previoulsly functionalized with haptoblobin antibodies [18].
134 C. Quero et al. / Chemico-Biological Interactions 192 (2011) 129–135
Fig. 5. Resolution of the signals from the ␣-1 chain after modification of the lysine
residues with O-methylisourea. The signal from the ␣-1 chains is shifted from m/z
9192 to m/z 9526 (␣-1s) and 9568 (␣-1f), resulting in a mass difference of 42 units
between these chains.
Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission from
Ref. [17].
3.3. Hp phenotype distribution
The distribution of Hp alleles varies greatly between human
populations, and there are important differences between ethnic
groups. The analysis of Hp distribution is thus a useful tool for
genetic population studies [24]. Numerous Hp subtype distribution
studies in different populations including small, ethnically well-
defined communities have been reported [25–27]. Surveys in Spain
showed a distribution of about 63% and 60% (HP2 ), 24% and 23%
(HP1S ) and 14% and 17% (HP1F ) for Barcelona (Mediterranean coast)
and Madrid (central area), respectively [28,29]. In the case of our
population, which corresponds to individuals of the central area of
Spain, these values were of 57% (HP2 ), 31% (HP1S ) and 12% (HP1F )
(Table 1). The observed deviation from the reported values is mainly
accounted for by the contribution of the case population where the
HP1S and HP1F are over- and underrepresented, respectively. The
comparison of phenotype distributions (Table 2) showed signifi-
cant differences between control and cases of the Hp2-1s and 2-1f
phenotypes (p = 0.03 and p = 1.461027, respectively). Nearly 70% of
the Hp2-1s individuals were cases, while only 20% with the Hp2-1f
phenotype were in this group. The OR for Hp2-1s in control versus
cases was 0.46 (p = 0.09, 95% CI = 0.184–1.158), while for Hp2-1f was
6.0 (p = 0.01, 95% CI = 1.223–29.348). The Hp2-2 phenotype showed
a similar proportion in case and controls, suggesting that the HP2
allele is not related to the susceptibility to the disease. The same
trend is observed for the Hp1s-1f phenotype, although in this case
Table 1
Allele distributions (%) of Hp in the studied population.
Allele Control TOS
HP1s 24.0 38.0
HP1f 17.7 6.5
HP2 58.3 55.6
Table 2
Phenotype distributions (%) of Hp in the studied population.
Phenotype Control TOS Total
1s-1s 4 9 13
Hp 1-1 1s-1f 6 5 11
1f-1f 1 0 1
2-1s 9 18 26
Hp 2-1
2-1f 9 2 11
Hp 2-2 2-2 19 20 38
this result would be consistent with an antagonistic effect between
the Hp1s and 1f chains.
These results suggest that the Hp1f-1f phenotype could confer
resistance to the disease.
4. Concluding remarks
We have developed a MALDI-TOF profiling method for the sub-
typing of Hp that outperforms the speed and efficiency of gel-based
procedures and makes no use of immunodetection procedures. The
analysis is carried out in two steps: determination of the presence
of the ␣-1 and ␣-2 chains, followed by subtyping of the ␣-1 chain
in the corresponding samples as ␣-1s and/or ␣-1f. The technique
is not affected by the presence of some known modified Hp chains
that can interfere with the detection in 2-DE gels and can be applied
for the large scale typing of Hp. We applied this method to the anal-
ysis of a group of serum samples from TOS patients confirming our
previous results obtained by 2-DE/MALDI-TOF from pools of sera
that indicated that the presence of the HP1s allele of Hp was a risk
factor for the TOS disease, while the HP1f was a protective factor.
Hp, which modulates many aspects of the acute-phase response, is
known to have important physiological functions, many of which
are related to its antioxidant properties.
Conflict of interest statement
The authors declare that there are no conflicts of interest.
Acknowledgements
This work was supported by WHO projects EU/03/016989 and
EU 05/025039.
References
[1] J.M. Tabuenca, Toxic allergic syndrome caused by ingestion of rapeseed oil
denatured with aniline, Lancet 2 (1981) 567–568.
[2] E. Gelpí, M. Posada de la Paz, B. Terracini, I. Abaitua, A. Gómez de la Cámara,
E.M. Kilbourne, C. Lahoz, B. Nemery, R.M. Philen, L. Soldevilla, T. Stanislaw,
(WHO/CISAT Scientific Committee for the Toxic Oil Syndrome), The Spanish
toxic oil syndrome 20 years after its onset: a multidisciplinary review of scien-
tific knowledge, Environ. Health Perspect. 110 (5) (2002) 457–464.
[3] R.E. Calaf, J. Pena, S. Paytubi, B.C. Blount, M. Posada de la Paz, E. Gelpí, J. Abian,
Determination of aniline derivatives in oils related to the toxic oil syndrome
by atmospheric pressure ionization-tandem mass spectrometry, Anal. Chem.
73 (2001) 3828–3837.
[4] H.J. Kaufman, R.J. Seidman, l-Tryptophan-associated eosinophilia-myalgia syn-
drome: perspective of a new illness, J. Rheum. Dis. Clin. North Am. 17 (1991)
427–441.
[5] M.G. Ladona, M. Izquierdo-Martinez, M.P. Posada de la Paz, R.C. de la Torre,
C. Ampurdanés, J. Segura, E.J. Sanz, Pharmacogenetics profile of xenobiotic
enzyme metabolism in survivors of the Spanish toxic oil syndrome, Environ.
Health Perspect. 109 (2001) 369–375.
[6] C. Quero, N. Colome, M.R. Prieto, M. Carrascal, M. Posada, E. Gelpí, J. Abian,
Determination of protein markers in human serum: analysis of protein expres-
sion in toxic oil syndrome studies, Proteomics 4 (2004) 303–315.
[7] H. Cleve, S. Gordon, B.H. Bowman, A.G. Bearn, Comparison of the tryptic pep-
Total tides and amino acid composition of the beta polypeptide chains of the three
31.4 common haptoglobin phenotypes, Am. J. Hum. Genet. 19 (1967) 713–721.
[8] G.E. Connell, O. Smithies, G.H. Dixon, Gene action in the human haptoglobins.
11.8
II. Isolation and physical characterization of alpha polypeptide chains, J. Mol.
56.9
Biol. 21 (1966) 225–229.
 C. Quero et al. / Chemico-Biological Interactions 192 (2011) 129–135
[9] V. del Pozo, B. de Andrés, S. Gallardo, B. Cárdaba, E. de Arruda-Chaves, M.I.
Cortegano, A. Jurado, P. Palomino, H. Oliva, B. Aguilera, M. Posada, C. Lahoz,
Cytokine mRNA expression in lung tissue from toxic oil syndrome patients: a
TH2 immunological mechanism, Toxicology 118 (1997) 61–70.
[10] B. Cárdaba, J. Ezendam, V. del Pozo, M. Izquierdo, C. Martín, I. Cortegano, E.
Aceituno, M. Rojo, I. Arrieta, P. Palomino, M. Posada, C. Lahoz, DR2 antigens are
associated with severity of disease in toxic oil syndrome (TOS), Tissue Antigens
55 (2000) 110–117.
[11] P.R. Jungblut, U. Zimny-Arndt, E. Zeindl-Eberhart, J. Stulik, K. Koupilova, K.P.
Pleibner, A. Otto, E.C. Muller, W. Sokolowska-Kohler, G. Grabher, G. Stoffler,
Proteomics in human disease: cancer, heart and infectious diseases, Elec-
trophoresis 20 (1999) 2100–2110.
[12] I. Noel-Georis, A. Bernard, P. Falmagne, R. Wattiez, Proteomics as the tool to
search for lung disease markers in bronchoalveolar lavage, Dis. Markers 17
(2001) 271–284.
[13] E.F. Petricoin, A.M. Ardekani, B.A. Hitt, P.J. Levine, V.A. Fusaro, S.M. Steinberg,
G.B. Mills, C. Simone, D.A. Fishman, E.C. Kohn, L.A. Liotta, Use of proteomic
patterns in serum to identify ovarian cancer, Lancet 359 (2002) 572–577.
[14] G.B. Mills, R.C. Bast Jr., S. Srivastava, Future for ovarian cancer screening:
novel markers from emerging technologies of transcriptional profiling and
proteomics, J. Natl. Cancer Inst. 93 (2001) 1437–1439.
[15] R.W. Putnam, The Plasma Proteins, Academic Press, New York, 1975.
[16] Y.Y. Wang, P. Cheng, D.W. Chan, A simple affinity spin tube filter method
for removing high-abundant common proteins or enriching low-abundant
biomarkers for serum proteomic analysis, Proteomics 3 (2003) 243–248.
[17] C. Rodriguez, C. Quero, A. Dominguez, M. Trigo, M. Posada de la Paz, E. Gelpí, J.
Abian, Proteotyping of human haptoglobin by MALDI-TOF profiling: phenotype
distribution in a population of toxic oil syndrome patients, Proteomics 6 (2006)
S272–S281.
[18] J. Abián Moñux, P. Eichhrorn, C. Rodríguez, Procedimiento para el tipado de
haptoglobina en suero humano, CSIC, Madrid, Spain, Patent ES 2310 431 B1
(2009).
135
[19] A. Shevchenko, M. Wilm, O. Vorm, M. Mann, Mass spectrometric sequencing
of proteins from silver-stained polyacrylamide gels, Anal. Chem. 68 (1996)
850–858.
[20] D.F. Hochstrasser, J.-D. Tissot, Clinical applications of high-resolution two
dimensional polyacrylamide gel electrophoresis, Adv. Electrophor. 6 (1993)
268–375.
[21] E. Darii, D. Lebeau, N. Papin, A.Y. Rubina, A. Stomakhin, J. Tost, S. Sauer, E. Savva-
teeva, E. Dementieva, A. Zasedatelev, A.A. Makarov, I.G. Gut, Quantification of
target proteins using hydrogel antibody arrays and MALDI time-of-flight mass
spectrometry, New Biotechnol. 25 (2009) 404–416.
[22] D. Finnskog, A. Ressine, T. Laurell, G. Marko-Varga, Integrated protein microchip
assay with dual fluorescent- and MALDI read-out, J. Proteome Res. 3 (2004)
988–994.
[23] J.R. Kimmel, Guanidination of protein, Methods Enzymol. 11 (1967)
584–589.
[24] M.R. Langlois, J.R. Delanghe, Biological and clinical significance of haptoglobin
polymorphism in humans, Clin. Chem. 42 (1996) 1589–1600.
[25] B. Teige, B. Olaisen, L. Pedersen, Haptoglobin subtypes in norway and a review
of Hp subtypes in various populations, Hum. Hered. 42 (1992) 93–106.
[26] A. Yano, Y. Yamamoto, S. Miyaishi, H. Ishizu, Haptoglobin genotyping by allele-
specific polymerase chain reaction amplification, Acta Med. Okayama 52 (1998)
173–181.
[27] J.R. Delanghe, K. Allcock, M. Langlois, L. Claeys, M. De Buyzere, Fast deter-
mination of haptoglobin phenotype and calculation of hemoglobin binding
capacity using high pressure gel permeation chromatography, Clin. Chim. Acta
291 (2000) 43–51.
[28] P. Moral, A.M. Panadero, Haptoglobin subtypes in Barcelona (Spain), Hum.
Hered. 33 (1983) 192–194.
[29] A. Alonso, G. Visedo, M. Sancho, J. Fernandez-Piqueras, Haptoglobin
subtyping by isoelectric focusing in miniaturized polyacrylamide gels rehy-
drated in the presence of 2-mercaptoethanol, Electrophoresis 11 (1990)
321–324.